[GRASS-user] advice on interpolation

Hi,

I have an ASTER GDEM (Latlong, 30 m resolution, from NASA web site) imagery. I want to produce 10m res DEM from it. Here is what I have done and what my problem is. I really appreciate suggestions.

  • I reprojected the imagery into a UTM location (using the ‘nearest’ method),
  • Using ‘g.region’, I set my region to match the original DEM
  • I set my computational extents from the display,
  • then, I sampled 50% of the original imagery (in r.random) and created a vector map (it took long time to do this and it has around 5 million points in it),
  • I set the resolution to 10 meters using ‘g.region’,
  • then, using ‘v.surf.rst’, I tried to make a new DEM -10m resol- from the sampled vector map and masking the operation with the original -30m- DEM. I do not change the default settings and I use a smoothing parameter of ‘.1’ and i use the “value” column from the vector map for approximation in the process.

I get a lot of error messages during interpolation, like;

G_ludcmp() failed! n=0
taking too long to find points for interpolation–please change the region to area where your points are. Continuing calculations…

I am not sure but may be my region settings is wrong (i.e., too wide) although I used the imagery to set them. I never waited for interpolation to run completely.

Thank you for suggestions!


BÜLENT ARIKAN
School of Human Evolution and Social Change
Arizona State University
Tempe - AZ
85287-2402

On Tue, 30 Mar 2010, Bulent Arikan wrote:

- Using 'g.region', I set my region to match the original DEM

Bulent,

   Do you use the '-a' (alignment) option on g.region? You'll often get
errors if an edge does not exactly line on the region's edge. This option
removes this source of potential errors.

Rich